Singapore’s Ten-Cent Bottle Scheme Could Turn Restaurant Waste Into a Cash-Recovery Job.
Merchant-owned deposits may support a shared collection and allocation service.
The Signal in brief
- Who it affects
Existing cleaners, building managers and participating restaurant clusters.
- Where
Singapore.
- Urgency
Measure by 24 November; avoid new vehicles until paid economics are verified.
Paid demand, route costs and material merchant deposit losses remain unverified.
Explore this signal
Emerging data
A used drink container can now hold money that belongs to a restaurant.
On 4 November, the Ministry of Sustainability and the Environment said about 3,000 outlets were in the Return Right F&B Scheme. Participating outlets absorb the ten-cent deposit for drinks consumed on the premises instead of passing it to dine-in customers. They retain the empty containers. This creates a business task: collect the right containers, return them and make sure the right outlet receives the money. It does not apply to every restaurant or every drink sale.
The machines and processing plant are already operating. More than 48 million containers had been collected by 27 September. Yet September reporting put the counting plant’s daily throughput at about 20 tonnes against 45 tonnes of capacity. On 9 November, Chinese-language reporting described a queue and an alleged machine problem at Tekka Market. These observations concern different dates and places. Together, they suggest that local collection, timing and convenience deserve testing; they do not prove a national machine shortage.
Our database’s 4 November report already identified refund handling, customer support and payment reconciliation as possible operational opportunities. The new connection is merchant-owned deposits plus several outlets sharing one collection route. We think an existing cleaner or building manager has a better starting position than a new operator buying vehicles: the person already handling the containers may add the task at lower cost.
Clustered container collection and deposit allocation could become a small recurring service where measured savings exceed its fee. Machine maintenance and field-service work is another route: TOMRA’s current Singapore technician listing advertises S$2,600–3,300 a month. TOMRA and Chye Thiam Maintenance already have a service relationship, so the market has established providers.
The arithmetic keeps this grounded. An illustrative 30,000 accepted containers per month returns S$3,000 to the outlets. A proposed three-cent service fee earns the provider S$900, taking 30% of the recovered deposit. If extra collection and administration cost S$600, the provider retains S$300 before overhead and tax. None of those service prices or costs is a verified customer contract. The offer works only if the outlets save more than S$900 in otherwise lost refunds and labour.
The immediate opening is a two-week measurement exercise with an existing outlet cluster. Compare the current collection arrangement, containers lost, labour time and actual refunds. A new fee is justified only after that comparison and a signed paid trial. Installing more machines or hiring a full-time collector comes later, if the measured economics support it.
The connection
A service fee only works if recovered money and labour savings exceed its cost.
The opportunity
Measure two weeks of containers, staff time and refunds before a written paid trial.
What could happen next
- Second-order consequence
Second order — outlets measure discarded cash. If participating restaurants track containers against purchases, some may discover avoidable deposit losses or expensive staff trips.
- Third-order consequence
Third order — nearby outlets combine the work. Enough verified losses could justify a shared collection route. Density matters: an existing cleaner may serve several outlets during an existing visit.
- Fourth-order consequence
Fourth order — records become part of the service. Shared collections require a reliable way to allocate accepted containers and refunds. Building managers could bundle that work with cleaning, provided the total fee beats current alternatives.
- Fifth-order consequence
Fifth order — reliable returns support the recycling system. Better collection could raise plant throughput and recurring maintenance demand. If existing arrangements already recover deposits cheaply, the outside service disappears before reaching this stage. The sequence is a forecast, not a count of awarded contracts.
What would change our view?
Disconfirmation — the fee must survive comparison. Reject the business if free or existing collection recovers the same money, the outlet already loses little, container rejection is high, or extra labour exceeds the fee. A busy return point can indicate successful participation as well as inconvenience.
Evidence and sources
- MSE: Return Right F&B SchemeView source
- NEA collection resultsView source
- CNA: counting and sorting plantView source
- TOMRA’s Singapore launchView source
Source records
Sources and timestamps. Singapore is the operating market unless stated otherwise. Source Publication; event Corroboration and confidence limits
S1 — MSE: Return Right F&B Scheme 4 Nov 2026, date only; parliamentary reply and current participation. Original government source for participating outlets and deposit treatment. Not a list of buyers for a new service.
S2 — NEA collection results Page dated 7 Oct; letter identified as published 9 Oct; measured 1 Apr–27 Sep. Official cumulative figures, not current November volumes or an independently audited return rate.
S3 — CNA: counting and sorting plant 21 Sep, 16:01 SGT; updated 22 Sep, 09:17; September throughput. Independent site reporting. The photograph is dated 16 Sep; the precise measurement day is unverified.
S4 — TOMRA’s Singapore launch 1 Apr, date only; scheme launch. Primary supplier account of machines and maintenance partner; deployment, not disclosed project profit.
S5 — 8world: Tekka return-machine queue 9 Nov, 21:08 SGT; recently posted local video, precise filming clock unverified. Chinese reporting based partly on a social-media witness. Alleged fault was not independently established.
S6 — MSE: unsuccessful refunds 4 Nov, date only; transaction experience as of October and portal rollout from 1 Oct. Official transaction measure. Does not measure machine uptime or merchant deposit recovery.
S7 — NEA operator guidance and scheme licence Licence page marked updated 4 Mar; guidance retrieved 10 Nov SGT; licence 1 Apr 2026–31 Mar 2033. Primary operating framework; current site-specific collection terms still need confirmation.
S8 — Singapore field-service vacancy Undated live listing, retrieved 10 Nov SGT. Primary salary and job description. One open role is not a hiring surge or completed hire.
S9 — TOMRA third-quarter report 23 Oct, date only; quarter ended 30 Sep; Norway/global. Original company accounts. Singapore results are grouped with other markets.
S10 — Director purchase and DNB holding reduction 30 Jul, 19:38:54; trade same day. DNB disclosure 30 Sep, 08:27:31; sale 29 Sep. Zones unstated; Norway. Formal market disclosures carried by MFN. Mixed ownership clues, unrelated to a specific Singapore contract.
S11 — TOMRA market quote 6 Nov 2026 closing price; Oslo historical data, rechecked 10 Nov SGT. NOK99.60 historical close; not a live entry price. The original trailing-valuation snapshot could not be reproduced.
S12 — OTR’s prior refund-handling note 4 Nov, 17:57:17 SGT report cutoff; editorial source verification recorded at 18:14:27 SGT. Original archived report re-read. It proposes refund-handling and reconciliation work as an untested synthesis, not a named contract or completed paid-service outcome.
S13 — CNA: existing bulk collection 16 Mar, 06:00 SGT; updated 08:49; pre-launch arrangements. Direct NEA/BCRS reply, reported independently. Current eligibility and charges must be reconfirmed.
S14 — TOMRA 2025 annual report Released 20 Mar 2026; year ended December 2025; Norway/global. Original report on the exchange. Historical group concentration, not Singapore or 2026 concentration.
Verified source links S1: https://www.mse.gov.sg/latest-news/written-reply-to-parliamentary-question-on-return-right-f-b-scheme---exemptions-from-bcrs/ S2: https://www.nea.gov.sg/media/readers-letters/index/beverage-container-return-scheme-encourages-recovery-for-clean-stream-recycling S3: https://www.channelnewsasia.com/singapore/beverage-container-return-scheme-counting-sorting-facility-6393946 S4: https://www.tomra.com/en-gb/reverse-vending/media-center/news/2026/singapore-deposit-return-scheme-launch S5: https://www.8world.com/singapore/tekka-market-bcrs-3335506 S6: https://www.mse.gov.sg/latest-news/oral-reply-to-parliamentary-question-on-failed-reverse-vending-machines--bcrs-/ S7: https://www.nea.gov.sg/our-services/waste-management/beverage-container-return-scheme/for-f-b-operators | https://www.nea.gov.sg/our-services/waste-management/beverage-container-return-scheme/the-scheme-operator--bcrs-ltd S8: https://jobs.smartrecruiters.com/Tomra/744000159057036-field-service-technician-singapore- S9: https://cdn.financialreports.eu/financialreports/media/filings/3775/2026/RNS/3775_rns_2026-10-23_969d543f-90e8-416d-9408-324237c5e6af.pdf S10: https://mfn.se/all/a/tomra-systems/tomra-mandatory-notification-of-trade-by-pdmr-4a91137a | https://www.mfn.se/ob/a/tomra-systems/tom-major-shareholder-disclosure-tomra-systems-37f53efa S11: https://stockanalysis.com/quote/osl/TOM/history/ S13: https://www.channelnewsasia.com/singapore/drink-can-beverage-return-deposit-restaurants-5988441 S14: https://live.euronext.com/sites/default/files/company_press_releases/attachments_oslo/2026/03/20/668955_TOMRA_Annual_Report_2025.pdf
Explore the research
Research context
Research checked through
10 November 2026, Singapore time. See source records for each publication and event date, retrieval limit and remaining gap.
20 intelligence layers
1. Money movement
Money movement — the deposit already exists. Participating outlets carry the refundable amount for dine-in drinks. The ten cents is the outlet’s recoverable money, not a collector’s sales revenue. A collector earns only the separately agreed fee.
2. Early purchasing
Early purchasing — infrastructure is deployed. TOMRA reported 350 machines at launch and an established maintenance partner. That is evidence of real infrastructure and service activity. Public values of those contracts and new merchant orders for the proposed collection service were not retrieved.
3. Supply bottlenecks
Supply bottlenecks — location can matter more than total capacity. September’s spare processing capacity and November’s local queue are compatible: returns can be uneven across locations and hours. They do not establish that spare plant capacity can solve a particular market’s access problem.
4. Hidden beneficiaries
Hidden beneficiaries — existing routes have an advantage. Cleaners, building managers and qualified repair providers could add collection, allocation or maintenance work. Chye Thiam Maintenance is already named in the deployed system. No new award to an independent entrant was verified.
5. Company financial exposure
Company financial exposure — TOMRA is broad exposure. Collection supplied €217 million of €366 million quarterly group revenue, with a 15.1% operating margin. Poland, Singapore and Romania together supplied €24 million. Cash was €67 million, interest-bearing debt €500 million, with leases separate. In 2025, no customer supplied 10% or more of group revenue. Singapore-only earnings remain unknown. TOM, Oslo, therefore provides broader exposure.
6. Unit economics
Unit economics — an explicit trial model. The story’s assumed S$900 fee less S$600 incremental cost leaves S$300 a month before overhead and tax. An assumed S$600 setup would have a two-month simple payback. This excludes a new vehicle and salaried collector. Count rejected containers, extra trips and payment delays before relying on it. No paid trial verifies these assumptions.
7. Insider and institutional activity
Insider and institutional activity — conflicting clues. Director Pierre Couderc bought 1,600 TOMRA shares at NOK103 in July. A DNB-managed fund later sold 107,928 shares, taking its holding to 4.96%. Neither establishes how Singapore will perform.
8. Hiring and organisational changes
Hiring and organisation — practical work is visible. TOMRA’s Singapore listing includes repair, cleaning and machine monitoring at the stated S$2,600–3,300 salary range. This supports a direct employment route. It does not establish labour scarcity or a subcontract available to anyone.
9. Government spending pipeline
Government pipeline — operating rules, not a new grant. BCRS Ltd’s licence runs from April 2026 to March 2033. The framework provides continuity. No new government award or available startup subsidy for this proposed business was verified; an earlier signup incentive must not be treated as still open.
10. Foreign-language early warnings
Foreign-language warning — inspect the reported location. The Chinese report identifies a specific place to check. One witness account cannot establish national reliability or persistent downtime.
11. Consumer behaviour before sales
Behaviour before sales — participation does not prove outsourcing demand. A merchant’s willingness to pay depends on its labour cost, lost deposits and current route. Observe those costs directly before offering a fee.
12. Market contradictions
Market contradiction — different measures tell different stories. More than 95% of reverse-vending transactions received immediate refunds as of October. That can coexist with a queue or an unsuccessful visit elsewhere. It is not a 95% uptime statistic, national collection rate or restaurant recovery rate.
13. Missing expected activity
Missing expected activity — some refunds need follow-up. A new portal addresses unsuccessful refunds, with the expected resolution time shortened to ten working days. That is a documented exception to immediate payment. The number of merchants losing material deposits is still unknown.
14. Commercialisation
Commercialisation — the process works at scale already. Machines, sorting and refund systems are in operation. A new provider does not need to invent a recycling system. It must establish permission, suitable collection terms and accurate accounting around an existing one.
15. Competitive weaknesses
Competitive weakness — bulk collection already exists. BCRS described collection for outlets accumulating at least 500 containers, including combined returns. Confirm today’s terms before charging for overlapping work. The possible opening is better coordination or lower handling cost, not access to an otherwise missing service.
16. Geographical spillovers
Geographical spillovers — overseas equipment creates local service work. Norwegian TOMRA’s Singapore deployment includes Chye Thiam Maintenance as service partner. This is a real cross-border link between equipment supply and local operations. Adapting the collection model elsewhere would still require checking that country’s deposit rules and route economics.
17. Historical OTR patterns
Historical OTR pattern — a six-day refinement, not a scored win. The 4 November database report connected unsuccessful refunds with customer support, payment reconciliation and trust. Its broader synthesis explicitly warned that no named vendor had won an order. By 10 November, this draft narrows the customer and economic test to participating outlets. No paid result or verified failure of the earlier proposal was retrieved.
18. Awareness and pricing
Awareness and pricing — the system is well known. The opening, if any, is an inefficient local route. TOMRA’s 6 November closing price was NOK99.60. The supplied trailing-earnings multiple could not be reproduced and is not used here. Singapore’s contribution is not separately disclosed, so the evidence does not establish that the shares underprice this opportunity.
20. Immediate action triggers
Immediate triggers — measure by 24 November. Use an existing outlet relationship to record two weeks of eligible containers, staff time and accepted refunds. Start only after a written paid scope and confirmed collection terms. Cap setup at the illustrative S$600 until a month of paid work verifies the margin. Stop if the outlet’s measured savings do not exceed the fee; capital for a new vehicle is not justified by this evidence.
Evidence position
Evidence position and remaining gaps
The ten-cent deposit belongs to the outlet; service fees and route costs are illustrative, and existing bulk collection must be compared. Paid demand, route costs and material merchant deposit losses remain unverified. Facts, proposed trial economics and conditional forecasts remain distinguished.
Signal history
Published changes
Updated the evidence and sources and expanded research.